simming spanish deck games in cvdata -- penetration problem

[rrwoods]

[heh]

So I've got a shiny new CVData and I'd like to sim some eight-deck Spanish 21. The penetration numbers don't make sense to me.

If I'm selecting six deck, everything is as expected; I select cutting off 72 cards, which is one and a half spanish decks, and the penetration shows as 4.5/6. But if I go to eight decks it shows as 6/8 (which is oddly enough the same percentage as it would be in six decks but not the right number of decks). Which number should I trust?

 

[QFIT]

[heh]

So I've got a shiny new CVData and I'd like to sim some eight-deck Spanish 21. The penetration numbers don't make sense to me.

If I'm selecting six deck, everything is as expected; I select cutting off 72 cards, which is one and a half spanish decks, and the penetration shows as 4.5/6. But if I go to eight decks it shows as 6/8 (which is oddly enough the same percentage as it would be in six decks but not the right number of decks). Which number should I trust?

Oddly, both. I need to explain this better. When you create a strategy based on the SP21 template, it is a six-deck game. If you wish to change it to an eight-deck game, you will need to change the modified deck table on the payoffs screen. Then, the numbers will make sense. Seems odd, but it is designed for max flexibility. This allows you to test, for example, SP21 wih an ace removed.

 

[rrwoods]

Oh wow, OK that makes perfect sense now that I see what you mean.

Perhaps put the modified shoe options on the setup screen along with the number of decks option? That way you can update the numbers in the shoe at the same time and the user will see it; if they want to cut specific cards it's right there.

 

[QFIT]

The goal is flex. It's rather a pain figuring out where and how to put the options.

 

[rrwoods]

OK I'm actually having a quite hard time figuring out how to sim Spanish 21 properly. I'd like to run a CVCX-style sim in order to optimize a betting ramp for various bankrolls, penetrations, etc, using the true-counted Hi-Lo system in Walker's book. A step-by-step of this would be really helpful.

 

[QFIT]

SP21 is beyond CVCX parallel sims. CVData must be used for complex games/strategies. But, you should be able to run a CVData SP21 sim and then call CVCX and see calcs for the specified settings.

 

[rrwoods]

So I tell CVData I'm using a betting strategy, but I'm actually lying; then after the sim I call CVCX to optimize my bets?

 

[QFIT]

So I tell CVData I'm using a betting strategy, but I'm actually lying; then after the sim I call CVCX to optimize my bets?

If you wish.

The CVCX simulator has no betting strategy at all when it is run. It applies betting after the sim. A highly unusual way of running sims, but allows for adjustments post-sim and optimal betting calcs. CVData requires an exact strategy before hand, as it supports extremely concept cover strategies that are impossible to apply post-sim. If you call CVCX from CVData, CVCX will ignore the CVData betting and apply either optimal betting, or whatever fixed bets per count you apply.

Again, the goal is max flexibility.

 

[rrwoods]

One more (last?) question:

I've made a strategy composed of all the Spanish 21 X-card strategies with the counts set to -1/1/1/1/1/1/0/0/0/-1 and IRCs of -4*decks. I've also set the counts for the "main" strategy that way. I'd expect to see the count distributions in Walker's book on the True Count Distributions view, but they seem to be centering around -2.5 rather than around -4. Is there anything I described that sounds obviously wrong, and if not, are there any gotcha's with counts when simming games with composite strategies?

 

[rrwoods]

OK it seems I had the true count divisors wrong :-/ However now it's centering around -4.3ish. I see 28% of hands at -4, 22% at -5, and 11% at -3.

 

[QFIT]

That would happen with flooring.

 

[rrwoods]

Aaaah, ok. I will have to look for that.